© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 4-RES-SM
    11. 10990
    Prev
    Next
    INDEX
    Explanations

    mentions of coffee and tea

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/4-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.4.hook_resid_post
    Hook Layer
    4
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    263
    +0.14
    0.8%
    306
    +0.14
    0.8%
    228
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    263
    +0.14
    0.09
    137
    +0.14
    0.08
    410
    +0.13
    0.05
    Negative Logits
    imization
    -2.36
    ifying
    -2.21
    ially
    -2.20
    ifiable
    -2.13
    ial
    -2.13
    ials
    -2.11
    ialize
    -2.09
    ile
    -2.08
    iating
    -2.02
    iam
    -1.99
    POSITIVE LOGITS
    maker
    2.06
    stick
    2.04
     Bear
    2.00
    ĻĤ
    1.90
     Tree
    1.83
     Bean
    1.79
    ¿½
    1.77
    cloth
    1.77
     maker
    1.76
    hurst
    1.76
    Activations Density 0.754%

    No Known Activations